Initial recovery trajectories among patients with hip fracture: a conceptual approach to exploring comparative effectiveness in postacute care

Abstract

Objective: To assess whether clusters of patients with hip fracture and with distinct initial recovery trajectories (IRT) could be identified by using practice-based evidence data and to examine the validity of these data.

Design: Analysis of multisite prospective observational cohort study database.

Setting: Eighteen skilled nursing and inpatient rehabilitation facilities.

Patients: Patients with hip fractures (N = 226) treated with joint replacement and admitted to skilled nursing or inpatient rehabilitation facilities, subset (n = 85), with telephone follow-up results approximately 8 months after rehabilitation discharge. Patients' ages were 76.8 ± 11.4 years; the majority were women (78%) and white (87%).

Methods: Measurements included medical severity by using the Comprehensive Severity Index and functional levels by using Functional Independence Measure (FIM). The IRT was calculated for each patient as the rate of change in function from the time of surgery to rehabilitation admission. We used cluster analysis to partition patients into subsets that shared common IRT scores. Validity was explored by comparing subgroups across patient characteristics and treatment patterns. Significance was defined as P ≤ .05.

Main outcomes measurements: IRT grouping, Comprehensive Severity Index, FIM, discharge location, living location at follow-up.

Results: We identified 3 patient clusters with differentiated IRT scores: group 1: 4.96 ± 2.45 (range, 0.4-8.6) FIM point change per day; group 2: 12.42 ± 2.51 (range, 8.9-17.0); group 3: 26.80 ± 13.78 (range, 17.5-70.0). Clinical group validation was established from statistically different Comprehensive Severity Index scores on admission; FIM scores at admission, discharge, and follow-up; and discharge and 8-month living settings. Calculation of IRT recovery curves by using FIM scores and associated time in days through logarithmic regression curves confirmed each group's IRT remained distinct through rehabilitation and follow-up.

Conclusions: The IRT concept appears to be valid in patients with hip fracture who were treated with hip arthroplasty, and may assist in evaluating and comparing the effectiveness of postacute rehabilitation services.